AMD Starts 6th Gen EPYC Venice 2nm CPU Production with TSMC Partnering for AI Processing and Verano Road Map Expansion
AMD has entered the production ramp for its next generation 6th Gen EPYC server CPUs codenamed Venice, manufactured using the 2nm process. In a joint statement from AMD and TSMC, the product has been touted as the first high performance computing product to move into production on a 2nm node. Production is currently focused in Taiwan but the chip will also be manufactured in TSMC s facility in Arizona, creating geographically diverse production capabilities for this chip generation.
2nm silicon production occurs as enterprises are being flooded by the intensive processing requirements of agentic AI. Within enterprise systems, the CPU typically acts as the orchestra, orchestrating traffic management, data transfer, storage access and security functions across the entire system.
"massive artificial intelligence workloads require enterprise server platforms capable of transforming innovation from lab to production rapidly."
AMD Chair and CEO Lisa Su stated the above while TSMC CEO C.C. Wei mentioned that:
"this partnership brings the power of both a leading foundry and innovative design to bear to create the future of enterprise computing."
To manage these intensive loads, AMD is coupling the 2nm process technology with integrated packaging technologies. For the Venice architecture, TSMC and AMD are incorporating custom TSMC technologies such as SoIC X and CoWoS L which allow AMD to stack dies close together, increasing overall power efficiency and decreasing latency for both Venice and AMD s larger server portfolio.
The 2nm EPYC server portfolio will be rapidly expanded by AMD, with the announcement of its follow on 6th Gen EPYC processor codenamed Verano. This follow on processor is designed to maximize performance dollar watt for power constrained cloud environments. Verano processors will come with innovative memory technologies integrated onto the processor, such as on package LPDDR, which is crucial for increasing bandwidth and performance to meet both the requirements for AI processing and cloud applications.
AMD is clearly prioritizing sustaining leadership in the server processor market through a clear focus on manufacturing and scaling. By coupling the raw silicon processing power of the 2nm node with modular design, high bandwidth memory solutions, and a focus on improving performance dollar watt, AMD aims to help businesses scale their computing capabilities.
